Improving Microwave Component Characterization using Wavelet Denoising Method

In this study, an efficient approach is presented to improve the characterization of various microwave components commonly used in communication and radar applications, such as antennas and power dividers. The components were initially simulated and then fabricated using the Computer Simulation Technology (CST) software. Vector Network Analyzer (VNA) measurements of the fabricated components were performed using a low-cost but noisy coaxial cable, and the measurement results were processed using a wavelet-based noise reduction method. For comparison purposes, the Haar and Daubechies-4 (DB4) wavelet functions were applied separately, and the results were examined. It was observed that the correlation and root mean square error between the ideal and measurement results improved in a positive direction with the noise reduction application. This approach provides significant cost and labor advantages, particularly when expensive elements such as gold and silver are used in coaxial cables that are physically free from noise. The experimental and numerical results show good agreement between the ideal simulation results and the filtered measurement results.